A Novel Microcontroller-Based System for the Wheel-Running Activity in Mice

Synthesis

1) The number of mice used for experiments was difficult to determine. Some of the experiments seem to have been run on a single mouse (i.e. chemogenetic activation of the perihabenular area). The number of mice should be clearly stated for all analysis, ideally written right on the relevant figures. Given the ease of the system, the authors should also consider expanding the number of mice in places with only one example mouse, as it will increase confidence in the robustness of the system for readers to see data from more than one mouse.

2) For the experiment in Figure 8 it would be helpful if the authors include an actogram or a summary figure of a control animal that received CNO but did not express the excitatory DREADD. Presumably this control animal would not exhibit a phase delay.

3) A hurdle for many neuroscientists in using open-source behavioral equipment is the analysis of raw data without commercial software. The authors should be commended on presenting python code to analyze wheel running data, yet this code is almost entirely not commented. It would be much more likely to be used if the authors could provide a brief description of each step or function. The authors should also include a more complete description of how to use the code than the included README.md, which is one fragmented sentence. They should also describe how they analyzed the data presented here in the text, and also discuss how this framework compares to other wheel running analyses.

4) The authors were commended for providing the Arduino code. However, it was also not well commented, which will reduce the possibility of users being able to implement or modify it. Please provide commented code, with headings to highlight each major part of the script.

5) The discussion/comparison of WRAQ with current technologies should be expanded to include wheel running as traditional home cage setups. The authors should describe how their technology compares to currently available commercial options.

6) Both reviewers noted that the grammar and language use in many placed was awkward and inhibited reading of the text. The authors should consider a thorough proof-reading to improve the language.

7) Some language should be revised to improve clarify. Specifically, reviewer 1 pointed out:

a. In the abstract, the authors should clarify that WRAQ doesn’t analyze the neural substrate regulating wheel running, it measures wheel running.

b. Line 7 refers to “the conventional cage system", yet there are many rodent caging systems widely in use and not one standard system. The authors should update this language to clarify that WRAQ likely fits in many homecages, with dimensions given so a reader can determine if it fits in their cages.

c. Line 31 refers to home cage activity as “the basis of analysis of behavioral phenotypes", which is not clear what is meant by this.

d. Line 46 refers to “genetically modified mice” as the model to be tested; WRAQ can be used to study wheel running in any mouse line.

e. And in line 47, the language should be changed to describe how researchers using this tool can modify inputs to the mouse condition, whether that’s chemogenetic, optogenetic, or something else in the environment.

8) Some of the figures could be combined into individual panels to create a more concise manuscript.

Minor points:

Figure 4E was truncated and could not be fully viewed.

Figures 1 and 2 and 5 do not seem to be mentioned in the body of the manuscript.

For Figure 3B, can the authors show average wheel running data over the light/dark cycle from all animals?

Please show an enlarged panel of c-fos staining (Fig 8C).

The BOM should be included as a figure in the paper, not as a supplemental excel file.

Author Response

Responses to the reviewers’ comments

1) The number of mice used for experiments was difficult to determine. Some of the experiments seem to have been run on a single mouse (i.e. chemogenetic activation of the perihabenular area). The number of mice should be clearly stated for all analyses, ideally written right on the relevant figures or included in the figure legends. Given the ease of the WRAQ system to collect data, the authors should consider expanding the number of mice in places with only one example mouse, as it will increase confidence in the robustness of the system for readers to see data from more than one mouse.

To address this point, we performed a novel experiment to examine the effect of chemogenetic activation of the suprachiasmatic nucleus (SCN)-subparaventricular zone (SPZ) instead of perihabenular area (PHb) on the wheel running activity in the revised manuscript. This is because we observed a variable effect of the chemogenetic activation of PHb on wheel running activity in the later analysis, probably because of the possible heterogeneity of PHb differentially targeted by viral vector across trials.

In a novel experiment, we measured the wheel-running activity of a mouse targeted by AAV8-hSyn-hM3D-mCherry to upregulate neuronal activity in the SCN-SPZ upon systemic administration of CNO (new Figure 6A and 6C). Studies showed that SCN-SPZ acted as a master clock and a region relaying the circadian information from the SCN to other brain regions, respectively (Ibuka and Kawamura, 1975; Lu et al., 2001). We measured the wheel-running activity of a mouse targeted by AAV8-hSyn-hM3D-mCherry to upregulate neuronal activity in the SCN-SPZ upon systemic administration of CNO (Figure 6A and 6C) (n = 5 mice). We observed a significant increase of c-Fos-positive cells in the SCN-SPZ following intraperitoneal injection of CNO (Figure 6C). Before CNO injection, the mouse exhibited free-running activity under DD (Figure 6A). Under the guidance of WRAQ-WiFi, injection of CNO at CT14 (red asterisks in Figure 6A) induced a significant shift in the onset of the active period lasting at least 7 days (Figure 6A).

We added this novel result as new Figure 6 and new Results section (p7, l230-l238 of the revised manuscript). We believe that this new result reproducible in the multiple mice will increase confidence in the robustness of the WRAQ system for the readers of our manuscript as the reviewer suggested.

References

Ibuka N, Kawamura H. Loss of circadian rhythm in sleep-wakefulness cycle in the rat by suprachiasmatic nucleus lesions. Brain Res. 1975 Oct 10;96(1):76-81.

Lu J, Zhang YH, Chou TC, Gaus SE, Elmquist JK, Shiromani P, Saper CB. Contrasting effects of ibotenate lesions of the paraventricular nucleus and subparaventricular zone on sleep-wake cycle and temperature regulation. J Neurosci. 2001 Jul 1;21(13):4864-74.

2) For the experiment in Figure 8 it would be helpful if the authors include an actogram or a summary figure of a control animal that received CNO but did not express the excitatory DREADD.

To address this point, we performed a control experiment in which the effect of CNO was administered systemically to the mice expressing GFP in SCN-SPZ. Results showed that intraperitoneal injection of CNO to the mice expressing GFP in the SCN-SPZ did not induce any of the effects on the onset of active period (Figure 6B) and c-Fos in SCN-SPZ (Figure 6D) significantly (n = 5 mice).

This new result was described in new Results section (p7, l238-240 of the revised manuscript).

3) A hurdle for many neuroscientists in using open-source behavioral equipment is the analysis of raw data without commercial software. The authors should be commended for including Python code to analyze wheel running data, yet this code is almost entirely not commented. It will be much more likely to be used by other scientists if the authors could provide a brief description of each step or function. The authors should also include a more complete description of how to use the code than the included README.md, which is one sentence that the reviewers and editor could not understand. A complete markdown file with screenshots of the analysis steps would be ideal. The authors should also describe how they analyzed the data presented here in the text, and also discuss how this framework compares to other wheel running analyses.

To address this point, we provided python program with comments to help the users to understand each step and function in the code. Furthermore, we provided a complete description of the workflow in processing the data from WRAQ in Readme.md file with screenshots of each analysis step. Since we will upload the extended data including the python program to the online repository such as github, extended data in this manuscript will be kept updated reflecting the feedback from the users in the future. Specifically, the raw data was converted into the file readable by ActogramJ, in which the data was supposed to start at zeitgeber time (ZT) 12, by custom-made python program (main_wraq2actj.py provided as new Extended Data, p4, l101-103 of the revised manuscript).

In comparison with the wheel running activity recording system available commercially, in which the data analysis could be done on the same platform using single software (e.g., ClockLab), data analysis in WRAQ system has an additional step to convert the raw data using pipeline python program to match the data format between data acquisition system (Adalogger and Ambient IoT server for WRAQ and WRAQ-WiFi, respectively) and analysis software (ActogramJ). This issue would be resolved by modifying the ActogramJ, which is also open-source software, to enable it to import raw data from WRAQ directly without conversion in the future.

This limitation was also added to the new Discussion section of the revised manuscript (p9, l284-l290).

4) The authors were also commended for providing the Arduino code for their running wheels. However, this code also had very few comments, which will reduce the possibility of users being able to implement or modify it. Please provide commented code, with headings to highlight each major part of the script. Please also include any library dependencies including version numbers.

In the revised manuscript, we intensively revised the Arduino code with additional comments which will help the users to implement or modify it as a part of new Extended Data. In addition, we revised a document which described how to install the dependent libraries with version information and programming environment in a step-by-step manner (“WRAQ_installation.md”). We believe that this will help the users to build their own environment more easily.

5) The discussion/comparison of WRAQ with current technologies should be expanded to include wheel running as traditional home cage setups. The authors should describe how their technology compares to currently available commercial options.

To address this, we rewrote the Discussion section and expanded to compare WRAQ with commercially available options.

Specifically, in comparison with the wheel running activity recording system available commercially, in which the data analysis could be done on the same platform using single software (e.g., ClockLab), data analysis in WRAQ system has an additional step to convert the raw data using pipeline python program to match the data format between data acquisition system (Adalogger and Ambient IoT server for WRAQ and WRAQ-WiFi, respectively) and analysis software (ActogramJ). This issue would be resolved by modifying the ActogramJ, which is also open-source software, to enable it to import raw data from WRAQ directly without conversion in the future.

We also noted in new Discussion section that the size of the WRAQ system (width 5 × depth 5 × height 3.5 inches) is smaller than representative low-profile running wheels (ENV-047, Med Associate Inc., Fairfax, VT, USA; width 6.1 × depth 6 × height 4 inches) used widely in neuroscience studies with mouse models. Since the latter low-profile mouse running wheel was used successfully in standard “shoebox” style individually ventilated cages (e.g., Model #9, Thoren Caging Systems, Inc., Hazleton, PA with width 7.70 × depth 12.17 × height 5.875 inches) (Beeler and Burghardt, 2021; Goh and Ladiges, 2015), it is reasonable to think that our WRAQ fits a wide variety of cage systems.

These paragraphs were added to the new Discussion section (p8, l264-l270 and p9, l284-l290 of the revised manuscript).

6) Both reviewers noted that the grammar and language use throughout was awkward and inhibited reading of the text. The authors should consider a thorough proof-reading to improve the language.

We asked Enago (www.enago.jp) for the English language review to improve the language in the main document as indicated the revised manuscript with highlights. We also proof-read the manuscript to improve the language and correct typos as well.

7) Some language should be revised to improve clarify. Specifically, reviewer 1 pointed out:

a. In the abstract, the authors should clarify that WRAQ doesn’t analyze the neural substrate regulating wheel running, it measures wheel running.

We rewrote this sentence as follows in the revised manuscript (p1, l14-l16).

"Together, these findings indicate that the WRAQ system is a novel and cost effective solution for the analysis of wheel-running activity in mice.”

b. Line 7 refers to “the conventional cage system", yet there are many rodent caging systems widely in use and not one standard system. The authors should update this language to clarify that WRAQ likely fits in many homecages, with dimensions given so a reader can determine if it fits in their cages.

According to the reviewer’s suggestion, we provided detailed information on dimension of the WRAQ and mouse cage we used in the current study in the new Methods section (p2, l61-p3, l65 and p3, l69-l71 of the revised manuscript).

The size of the WRAQ system (width 5 × depth 5 × height 3.5 inches) is smaller than representative low-profile running wheels (ENV-047, Med Associate Inc., Fairfax, VT, USA; width 6.1 × depth 6 × height 4 inches) used widely in neuroscience studies with mouse models. Since the latter low-profile mouse running wheel was used successfully in standard “shoebox” style individually ventilated cages (e.g., Model #9, Thoren Caging Systems, Inc., Hazleton, PA with width 7.70 × depth 12.17 × height 5.875 inches) (Beeler and Burghardt, 2021; Goh and Ladiges, 2015), it is reasonable to think that our WRAQ fits a wide variety of cage systems.

To address the reviewer’s point, we added this to new Discussion to help readers to determine whether WRAQ is applicable to their environment (p8, l264-l270 of the revised manuscript).

References

Goh J, Ladiges W. Voluntary Wheel Running in Mice. Curr Protoc Mouse Biol. 2015 Dec 2;5(4):283-290.

Beeler JA, Burghardt NS. Activity-based Anorexia for Modeling Vulnerability and Resilience in Mice. Bio Protoc. 2021 May 5;11(9):e4009.

c. Line 31 refers to home cage activity as “the basis of analysis of behavioral phenotypes", which is not clear what is meant by this.

We addressed this by rephrasing “the basis of analysis of behavioral phenotypes” with “a basic phenotype” in the revised manuscript (p2, l30).

d. Line 46 refers to “genetically modified mice” as the model to be tested; WRAQ can be used to study wheel running in any mouse line.

To address the reviewers’ comment, we replaced “genetically modified mice” to “mice” in the revised manuscript (p2, l43).

e. And in line 47, the language should be changed to describe how researchers using this tool can modify inputs to the mouse condition, whether that’s chemogenetic, optogenetic, or something else in the environment.

To address the reviewers’ comment, we rewrote this sentence as below to specify how researchers modify inputs to the mouse condition in the revised manuscript (p2, l44-l46).

"Streaming behavioral and environmental signals online would allow researchers to analyze the data in real-time and manipulate the ongoing neural activity at specific circadian times (CT) using genetic and pharmacological interventions.”

8) Some of the figures could be combined into individual panels to create a more concise manuscript.

According to the reviewer’s suggestion, we combined previous Figures 2 and 3 into new Figure 2 in the revised manuscript. We also combined previous Figures 4 and 6 into new Figure 3. This made our manuscript be more concise than before.

Minor points:

1. Figure 4E was truncated and could not be fully viewed. Please fix.

We fixed this problem and presented the entire panel for previous Figure 4E as new Figure 3E in the revised manuscript.

2. Figures 1 and 2 and 5 do not seem to be mentioned in the body of the manuscript. Please address.

We addressed this by referring to the previous Figures 1, 2 and 5 in the new Materials and Methods (p3, l71; p3, l73; p3, l78; p3, l80; p3, l82; p3, l84; p3, l92; p4, l101 of the revised manuscript) and new Results sections (p6, l201 of the revised manuscript).

3. Please show an enlarged panel of c-fos staining (Fig 8C).

In the revised manuscript, we added a magnified view of the panel of c-Fos staining as an inset in new Figure 6C to address this point.